bmp receptor ia is required in the mammalian embryo for endodermal morphogenesis and ectodermal...
TRANSCRIPT
www.elsevier.com/locate/ydbio
Developmental Biology 270 (2004) 47–63
BMP receptor IA is required in the mammalian embryo for endodermal
morphogenesis and ectodermal patterning
Shannon Davis,a Shigeto Miura,b Christin Hill,a Yuji Mishina,b and John Klingensmitha,*
aDepartment of Cell Biology, Duke University Medical Center, Durham NC 27710, USAbLaboratory of Reproductive and Developmental Toxicology, National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709, USA
Received for publication 16 September 2003, received 20 January 2004, accepted 27 January 2004
Available online 27 March 2004
Abstract
BMPRIA is a receptor for bone morphogenetic proteins with high affinity for BMP2 and BMP4. Mouse embryos lacking Bmpr1a fail to
gastrulate, complicating studies on the requirements for BMP signaling in germ layer development. Recent work shows that BMP4 produced
in extraembryonic tissues initiates gastrulation. Here we use a conditional allele of Bmpr1a to remove BMPRIA only in the epiblast, which
gives rise to all embryonic tissues. Resulting embryos are mosaics composed primarily of cells homozygous null for Bmpr1a, interspersed
with heterozygous cells. Although mesoderm and endoderm do not form in Bmpr1a null embryos, these tissues are present in the mosaics and
are populated with mutant cells. Thus, BMPRIA signaling in the epiblast does not restrict cells to or from any of the germ layers. Cells lacking
Bmpr1a also contribute to surface ectoderm; however, from the hindbrain forward, little surface ectoderm forms and the forebrain is enlarged
and convoluted. Prechordal plate, early definitive endoderm, and anterior visceral endoderm appear to be expanded, likely due to defective
morphogenesis. These data suggest that the enlarged forebrain is caused in part by increased exposure of the ectoderm to signaling sources
that promote anterior neural fate. Our results reveal critical roles for BMP signaling in endodermal morphogenesis and ectodermal patterning.
D 2004 Elsevier Inc. All rights reserved.
Keywords: BMPRIA; BMP signaling; Epiblast; Mouse; Forebrain; Ectoderm; Endoderm; Anterior visceral endoderm
Introduction Recent work indicates that BMP signals play many roles
Upon their formation at gastrulation, the primary germ
layers must undergo regionalization and morphogenesis to
establish the body plan. The mechanisms mediating these
early processes remain obscure. However, they clearly
depend on inductive interactions between the tissues of
the early gastrula. In vertebrates, the origin of many of the
cues for the initial regionalization of the germ layers is the
gastrula organizer and its derivatives, including axial mes-
endoderm and prechordal plate (PrCP) (Harland and Ger-
hart, 1997; Shimamura and Rubenstein, 1997). In mammals,
increasing evidence indicates that extraembryonic tissues
also play a critical role in establishing the body plan. For
instance, the anterior visceral endoderm (AVE) acts syner-
gistically with the gastrula organizer to initiate development
of the forebrain (Tam and Steiner, 1999).
0012-1606/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.ydbio.2004.01.048
* Corresponding author. Department of Cell Biology, Duke University
Medical Center, Box 3709, Durham NC 27710. Fax: +1-919-684-5481.
E-mail address: [email protected] (J. Klingensmith).
in the establishment of the vertebrate body plan, particularly
signaling by BMP4 and BMP2. These closely related
ligands utilize a common signal transduction cascade. In
vertebrates, BMP2 and BMP4 have been shown to utilize
either of two type I receptors, BMPRIA or BMPRIB, and a
type II receptor, BMPRII (Massague and Chen, 2000;
Miyazono et al., 2001). Upon ligand binding, the receptor
complex initiates a signal transduction cascade. We will
refer to this pathway as the BMP2/4 signaling pathway.
Each of the three germ layers appears to depend on
BMP2/4 signaling for its early development. In the naive
Xenopus ectoderm, BMP activity represses neural fate and
promotes surface ectoderm development (Weinstein and
Hemmati-Brivanlou, 1999). Additional evidence suggests
that BMP2/4 signaling inhibits anterior neural gene expres-
sion (Glinka et al., 1997; Hartley et al., 2001). BMP2/4
signaling promotes the generation of neural crest cells at the
neural – surface ectoderm boundary in chick and frog
(Knecht and Bronner-Fraser, 2002; Mayor and Aybar,
2001). In the mesoderm, BMP2/4 signaling promotes ven-
S. Davis et al. / Developmental Biology 270 (2004) 47–6348
tral identity in nascent tissue and limits the spatial extent of
organizer gene expression (Fainsod et al., 1994); later it
induces a cardiac fate in intermediate mesoderm (Schulth-
eiss et al., 1997). Finally, in endoderm, ectopic BMP2/4
signaling represses foregut markers, suggesting a role in
anterior–posterior patterning of the endoderm (Tiso et al.,
2002). Underscoring the importance of regulating BMP
signals after gastrulation commences, mouse embryos lack-
ing the BMP2/4 antagonists Chordin and Noggin show
defects in all three embryonic axes (Bachiller et al., 2000).
This implies that many developmental processes are sensi-
tive to levels of BMP2/4 signaling upon establishment of the
germ layers.
In mammals, a major obstacle in studying potential
requirements for BMP2/4 signaling in germ layer develop-
ment is the essential role it performs in initiating gastrula-
tion itself. Mouse embryos lacking BMP4 (Winnier et al.,
1995), BMPRIA (Mishina et al., 1995), or BMPRII (Beppu
et al., 2000) exhibit a block to gastrulation, such that
mesoderm does not form. However, both of these receptors
are expressed ubiquitously in post-gastrulation embryos
(Beppu et al., 2000; Dewulf et al., 1995) and are thus in
place to mediate BMP signaling during germ layer devel-
opment. Bmpr1b is not expressed in the gastrula (Dewulf et
al., 1995), and mice lacking this gene show no early defects
(Yi et al., 2000). Thus, BMPRIB is unlikely to play a major
role in generating the body plan, whereas BMPRIA is
probably a critical mediator of BMP2/4 signaling in this
process.
We have used embryonic stem cell chimeras and
tissue-specific gene ablations to determine the require-
ments for Bmpr1a in gastrulation and germ layer devel-
opment. We find that it is required in extraembryonic
tissues for primitive streak and mesoderm formation, but
is required in the epiblast for patterning and morphogen-
esis of the mesoderm and its derivatives (S.M., S.D., J.K.,
and Y.M., manuscript in preparation). Here we focus on
the roles of epiblast BMPRIA signaling in the develop-
ment of the ectoderm and endoderm. We used a condi-
tional null allele of Bmpr1a (Mishina et al., 2002) to
create embryos in which extraembryonic tissues are wild
type, but epiblast tissues are primarily composed of
mutant cells. These embryos gastrulate to form all three
germ layers, but exhibit severe problems in morphogen-
esis. The anterior neural ectoderm is dramatically enlarged
and convoluted, particularly the forebrain, at the expense
of surface ectoderm.
Materials and methods
Mice
Mox2-Cre heterozygotes (B6.129S4-Meox2tm1(cre)Sor)
(Tallquist and Soriano, 2000) were mated to Bmpr1a null
heterozygotes (B6.129S7-Bmpr1atm1Bhr) (Mishina et al.,
1995). Mox2cre/+; Bmpr1anull/+ progenies were mated to
homozygotes for a conditional allele Bmpr1aflox (129S7-
Bmpr1atm2Bhr) (Mishina et al., 2002), resulting in the
embryonic lethal genotype Mox2cre/+; Bmpr1anull/flox. Em-
bryos were generated by timed matings (Hogan et al., 1994)
and staged as described (Downs and Davies, 1993; Kaufman,
1992). Embryos and mice were genotyped by PCR (Mishina
et al., 2002) using yolk sac DNA or tail DNA, respectively
(Hogan et al., 1994). R26R (129S4-Gt(ROSA)26Sortm1Sor)
(Soriano, 1999) was crossed into the Bmpr1aflox line to
produce mice of the genotype Bmpr1aflox/flox; R26R/R26R.
These were mated to Mox2cre/+; Bmpr1a null/+ mice to
assay for recombination in wild-type and Bmpr1a mosaic
embryos.
Gene expression, proliferation, and histological sectioning
Whole mount in situ hybridization was preformed as
described previously (Belo et al., 1997). Probes used
were Ap2a (Mitchell et al., 1991), Foxa1, Foxa2 (Sasaki
and Hogan, 1993), Gbx2 (Wassarman et al., 1997), Gsc
(Camus et al., 2000), HesX1 (Thomas et al., 1995),
Krox20 (Swiatek and Gridley, 1993), Cer1 (Pearce et
al., 1999), mDkk1 (Glinka et al., 1998), Msx1 (Mackenzie
et al., 1991), Netrin-1 (Kennedy et al., 1994), Noggin
(McMahon et al., 1998), Otx2 (Ang et al., 1994), Shh
(Echelard et al., 1993), Six3 (Oliver et al., 1995), Sox2
(Wood and Episkopou, 1999), Sox10 (Kuhlbrodt et al.,
1998), T (Wilkinson et al., 1990), Wnt1 (Wilkinson et al.,
1987), and Wnt6 (Gavin et al., 1990). For AFP, a 906-bp
fragment encoding exons 3–9 of AFP (a gift from SM
Tilghman) was cloned into pBluescript (Stratagene). For
two-probe in situ hybridization, embryos were concur-
rently incubated with one probe labeled with digoxigenin-
UTP and a second probe labeled with fluorescein-UTP.
The first staining reaction was performed in NBT/BCIP,
after which the alkaline phosphatase was inhibited by a
5-min wash in 0.1 M Glycine, pH 2.2. The second color
reaction was preformed using 350 Ag/ml BCIP. Three or
more mutant embryos were examined for each probe. For
most tissues, patterning was assessed using multiple
probes.
Expression LacZ from the R26R locus was detected by
X-gal staining of embryos, followed by sectioning when
necessary (Hogan et al., 1994). To quantify recombination,
stained (recombined) cells and total cells were counted in a
defined area of tissue over three near adjacent sections. Cell
proliferation was assayed using an antibody against phos-
phorylated histone H3 (Upstate Biotechnology) as described
previously (Anderson et al., 2002).
Area and density measurements
Eight adjacent sections from three different axial levels
were used to measure the area of neural, mesoderm, and
surface ectoderm in three wild-type and three Bmpr1a
S. Davis et al. / Developmental Biology 270 (2004) 47–63 49
mosaic embryos using NIH Image. Measurements were
pooled at each axial level for a given tissue in each
embryo. For mesoderm density, sections from two wild-
type and two Bmpr1a mosaic embryos were stained with
DAPI (Roche). The total number of cells in a defined area
was determined from each of four sections in each embryo.
Wild-type and mosaic numbers were compared using a
paired T test.
Results
Ablation of Bmpr1a specifically in the epiblast
We designed an experimental approach to leave extra-
embryonic activity intact while reducing BMPRIA signal-
ing in embryonic tissues from before gastrulation. We used
a gene-targeting strategy in which the receptor is ablated
only in the epiblast. We employed a conditional null allele
of Bmpr1a (Mishina et al., 2002) in which an essential
exon is flanked by LoxP sites, the target sequences for Cre
recombinase. Mox2-Cre (MORE) expression begins at E5.5,
just before gastrulation, specifically in the epiblast (Tall-
quist and Soriano, 2000). Because the epiblast gives rise to
all the cells of the embryo proper, this approach yields a
conceptus deficient for Bmpr1a in embryonic tissues, but
wild-type in extra-embryonic tissues, before germ layer
formation.
In preparation for ablating Bmpr1a in the epiblast, we
examined the recombination activity of Mox2-Cre in our
experimental system. Mox2-Cre has been reported to be
mosaic in activity in some studies (Hayashi et al., 2002)
but not others (Wu et al., 2003). We crossed Mox2-Cre
mice with mice carrying the R26R Cre reporter locus
(Soriano, 1999), which expresses h-galactosidase (h-gal)in any cell in which Cre is active. At gastrulation through
organogenesis stages (embryonic day (E)7.0–9.5), activity
was observed in tissues of epiblast origin, but not in
tissues of trophoblast origin (Figs. 1A and E). However,
recombination was incomplete (Figs. 1B–D). Histological
sections revealed the highest level of recombination was
approximately 90% of cells (Fig. 1C). Most embryos are
primarily composed of mutant cells: of 38 embryos
assayed for recombination activity in whole-mount, 29
(76%) had well above 50% recombination. No specific
tissue appeared to have a higher or lower degree of
recombination, and labeled cells were distributed evenly
among all tissues.
As an independent means of determining whether recom-
bination byMox2-Cre is sufficient to recombine all Bmpr1a-flox alleles in epiblast derivatives, we used polymerase chain
reaction to amplify recombined and unrecombined Bmpr1a
conditional alleles in embryonic tissue from matings be-
tween Mox2-Cre; Bmprnull and Bmpr1aflox/flox parents. As
expected, the recombined Bmpr1aflox allele was only found
in embryos that carried Cre; however, all Cre-positive
embryos carried both recombined and unrecombined
Bmpr1aflox alleles (data not shown). In effect, then,
Mox2-Cre enables us to produce mosaic embryos com-
posed of mutant cells interspersed with wild-type (hetero-
zygous) cells, while trophoblast-derived (extraembryonic)
tissues remain wild type. We henceforth refer to these
Mox2-Cre; Bmpr1aflox/null conceptuses as Bmpr1a mosaic
embryos.
BMPRIA is required in the epiblast for multiple aspects of
germ layer development
All embryos mosaic for Bmpr1a activity gastrulate and
form a primitive streak, but exhibit severe and characteristic
defects in the development of all three germ layers. Pheno-
typic mutants can be identified from E7.5 by their dysmor-
phic shape (Fig. 2C). The most conspicuous defect is that
the anterior end is severely distorted, with an expanded,
convoluted ectoderm (Fig. 2D). The posterior is less affect-
ed, and an allantois forms but never fuses with the chorion.
Mosaic embryos fail to turn and maintain a cup-shaped
appearance, surviving until E10.5–11.0 (Table 1). They
form no heart and most likely die due to a lack of blood
circulation (data not shown).
Given the variability of recombination from Mox2-Cre,
we expected a wide range of phenotypes; however, we
found that the defects were quite consistent. We scored
embryos based on the appearance of the anterior end, the
most variable aspect of the phenotype. The vast majority of
mosaic embryos (92%; n = 74) had severe convolutions
such that discrete headfolds could not be recognized (Fig.
3A). A few (6%; n = 5) had recognizable but highly
dysmorphic headfolds (Fig. 3B). Very rarely (2%; n = 2),
mosaics had symmetric headfolds, but like the other classes
had stereotypic morphogenetic defects, including a failure in
ventral closure (Fig. 3C).
We used the recombination reporter R26R to analyze
recombination in Bmpr1a mosaic embryos. In these em-
bryos, Mox2-Cre catalyzes recombination simultaneously
of the Bmpr1aflox and R26R loci, in both cases leading to
the deletion of a small segment of DNA. Surprisingly,
embryos with widely different levels of recombination had
remarkably similar phenotypes. Embryos with nearly 90%
recombination looked very similar to embryos with less
than 50%, and would be scored as the most severe
phenotype (compare Figs. 3D–F). These results imply that
BMPRIA signaling controls a cell non-autonomous mech-
anism of embryonic morphogenesis.
Cells lacking BMPRIA populate definitive endoderm,
mesoderm, and ectoderm
We considered that the consistency of the mutant phe-
notype might be the result of mutant cells being dispropor-
tionately represented in a key tissue. If this tissue was in turn
the source of inductive signals controlling the patterning or
S. Davis et al. / Developmental Biology 270 (2004) 47–6350
growth of other tissues, one might expect to see a fairly
consistent phenotype regardless of the overall degree of
recombination. However, in examining many E7.0–8.5
Fig. 1.
Fig. 2.
mosaics in which mutant cells were marked, we found no
specific tissue or structure that had overtly more or fewer
mutant cells than other tissues in a given intact embryo. This
Table 1
Occurrence of Bmpr1a mosaic embryos at various embryonic stages
Age Number
of mutants
Total
embryos
Percentage
of mutant
E7.5 25 99 25
E8.5 116 493 24
E9.5 56 202 28
E10.5 8 54 15
E13.5 0 12 0
E14.5 0 8 0
E17.5 0 5 0
P0 0 15 0
Expected Mendelian ratios are observed before E10.5. After E10.5, Bmpr1a
mosaic embryos are not recovered.
S. Davis et al. / Developmental Biology 270 (2004) 47–63 51
was the case regardless of whether the embryo had a
relatively high or low level of recombined (mutant) cells.
To investigate this further, we quantified recombined cells in
a pair of sectioned embryos, Bmpr1a mosaic and wild-type,
that both appeared to be about 50% chimeric based on the
level of stained cells in whole-mount. The Bmpr1a mosaic
had the characteristic ‘‘strong’’ phenotype shown in Figs.
3D–F. No specific tissue or germ layer had disproportionate
levels of marked cells in either embryo (Figs. 3G and H;
data not shown). Mutant cells contributed at normal levels to
definitive endoderm and mesoderm, though these tissue
layers do not form in Bmpr1a null homozygotes due to
the lack of gastrulation (Mishina et al., 1995). Note that the
endodermal layer in these mosaic embryos must be defin-
itive endoderm-derived from the epiblast rather than tro-
phoblast—because it contains labeled cells reflecting the
epiblast-specific expression of Mox2-Cre. In short,
BMPR1A signaling in the epiblast does not restrict cells
to or from any of the germ layers or major tissues in the
early embryo. This indicates that BMPRIA is not required
for mesoderm or definitive endoderm cell identity or germ
layer formation.
Ectodermal domains are abnormal in embryos mosaic for
BMPRIA function
A striking defect of Bmpr1a mosaic embryos is that
the anterior of the embryo appears enlarged, with a
Fig. 1. Tissue distribution of Mox2-Cre activity. Recombined cells are marked by
locus. (A–B) Wild-type E8.25 embryos, lateral view. (A) Although this embryo ap
unrecombined cells. The extraembryonic yolk-sac shows much less staining becaus
of section in panel C. (B) This embryo exhibits a much lower level of recombi
counterstained section through A. Note unlabeled (pink) cells in all tissues. Inset b
unrecombined (pink) cells in ectodermal domains. (E) Extraembryonic tissues e
derivatives (arrowhead), but only unrecombined cells in trophoblast derivatives (
Fig. 2. Morphological defects of Bmpr1amosaic embryos at the end of gastrulation
(A) Wild-type E7.5 embryo, lateral view. (B) Sagittal section of the embryo shown
(am), and the allantois (al). In the embryo proper, the embryonic ectoderm (ec), m
embryo at E7.5. It has a broader, shorter profile than wild-type. (D) Sagittal sectio
including the allantois, amnion, and chorion. The most conspicuous defect is tha
portion of the embryo. The anterior mesoderm appears less organized.
deeply convoluted ectodermal layer. To characterize the
nature of ectodermal domains in mosaics, we assayed
expression of neural and surface ectoderm markers. Sox2
shows pan-neural expression in early mouse embryos
(Wood and Episkopou, 1999), being expressed throughout
the columnar epithelium of the neural tube (Figs. 4A–C).
In Bmpr1a mosaics, the deep ectodermal folds at the
anterior are composed of thick, columnar epithelium that
expresses Sox2 (Figs. 4D and E). Posterior sections have
a discrete neural groove, though the Sox2-expressing
columnar epithelium is not formed into a neural tube.
This epithelium merges laterally with low, cuboidal tissue
that does not label with Sox2, suggesting it is surface
ectoderm (Fig. 4F). We conclude that the columnar
epithelium in mosaics is neural ectoderm, relatively nor-
mal in morphology in the posterior but highly convoluted
and broad in the anterior.
Surface ectoderm is labeled specifically by Wnt6 ex-
pression (Gavin et al., 1990) and appears in wild-type
embryos as thin epithelium adjoining the neural plate
(Figs. 4G–I). Sections through the anterior of Bmpr1a
mosaic embryos lack Wnt6 expression in the ectoderm
(though it is expressed in the amnion), consistent with the
lack of thin, cuboidal epithelium in the anterior ectoderm
(Figs. 4J and K). In contrast, posterior sections show Wnt6
expression in the thin ectoderm lateral to the columnar
neuroepithelium (Fig. 4L). Histological data also indicate
that anterior sections have broad, thick (neural) ectoderm
and little if any thin (surface) ectoderm (Fig. 4M), while
posterior sections contain medial thick, columnar (neural)
ectoderm flanked by thin (surface) ectoderm (Fig. 4N), as
in wild type. Together, these results demonstrate that the
anterior ectoderm of Bmpr1a mosaics is virtually all
neural, while more caudal levels have both surface and
neural ectoderm.
BMPRIA is required for anterior surface ectoderm but not
for the surface ectoderm fate
Recent work in Xenopus suggests naive ectodermal cells
that undergo little BMP2/4 signaling are unable to become
surface ectoderm and become neural by default (Weinstein
h-galactosidase staining, demonstrating recombination at the R26R reporter
pears to be entirely marked in whole-mount view, sections reveal significant
e all but the yolk-sac mesoderm is of trophoblast origin. Line indicates level
nation, with a region of few marked cells (arrow). (C) Transverse, eosin-
oxes show areas shown in panels D and E. (D) Arrows indicate examples of
xhibit both recombined (blue) and unrecombined (pink) cells in epiblast
arrowhead). Scale bars = 0.25 mm, except D and E = 0.025 mm.
. In all panels, anterior is the left, posterior is to the right, and proximal is up.
in panel A. Extraembryonic structures include the chorion (ch), the amnion
esoderm (me), and endoderm (en) layers are indicated. (C) Bmpr1a mosaic
n of the embryo shown in panel C. Extraembryonic structures are present,
t the embryonic ectoderm is highly convoluted, particularly in the anterior
Fig. 3. Spectrum of phenotypes and recombination in Bmpr1amosaic embryos. (A) 91% of Bmpr1a mosaic embryos have severe distortions of the anterior end
(arrow). (B) 6% of Bmpr1a mosaic embryos have recognizable headfolds (arrow), although these structures are severely distorted. (C) 3% of Bmpr1a mosaic
embryos have symmetrical headfolds (arrow), but have a failure in ventral morphogenesis and other defects that characterize all Bmpr1a mosaics. (D–F)
Bmpr1a mosaic E8.5 embryos with varying degrees of recombination at the R26R reporter locus. Phenotypes are virtually identical through the levels of
recombination vary from approximately 90% (D) to less than 50% (F). (G and H) Quantitation of recombination in tissues of wild-type and mutant mosaic
embryos that appeared about 50% recombined at the R26R locus in whole-mount. (G) Percentage of recombined cells in a Mox2cre/+; RosaR26R/+ (wild-type)
embryo in embryonic mesoderm, neural ectoderm, and embryonic endoderm. (H) Levels of recombination in a Mox2cre/+; Bmpr1aflox/null; RosaR26R/+ embryo
in same three tissues. Error bars are FSEM. Scale bars = 0.5 mm.
S. Davis et al. / Developmental Biology 270 (2004) 47–6352
and Hemmati-Brivanlou, 1999). Although we observed
caudal surface ectoderm in Bmpr1a mosaics, the chimeric
nature of these embryos raises the possibility that this tissue
is entirely wild type, that is, it is possible that Bmpr1a null
cells are unable to participate in the formation of surface
ectoderm. Using R26R as a reporter of recombination in
Bmpr1a mosaic embryos, we assessed the genotypic com-
position of the surface ectoderm. We saw no exclusion or
abnormal contribution of recombined cells in the surface
ectoderm (Figs. 4P and Q); moreover, the distributions of
recombined cells were similar between surface and neural
ectoderm domains, and were consistent with distributions in
other germ layers (see also Figs. 3G and H). Thus, despite
the paucity of anterior surface ectoderm in Bmpr1a mosaics,
these data indicate that Bmpr1a is not required for surface
ectoderm cell fate per se.
Precursor tissues of the head are enlarged in Bmpr1a
mosaic embryos
The convolutions of the anterior neural epithelium in
mosaic embryos could result from an expansion of neural
tissue; however, they might also be explained by a defi-
ciency in head mesenchyme, which could cause the over-
lying neural epithelium to wrinkle without the support of
underlying mesoderm. Given that molecular marker analysis
confirmed that the thick, columnar ectoderm is indeed
neural, and that the contiguous cuboidal epithelium is
surface ectoderm, we used histology to identify tissues
and determined their surface area in serial sections. Relative
to wild-type tissues, the anterior neurectoderm of mosaic
embryos is indeed expanded, and there is a corresponding
expansion of underlying mesenchyme (Fig. 4O). The den-
sity of mesoderm cells did not differ significantly between
wild-type and mosaic embryos (P value = 0.35).
The enlarged neurectoderm of Bmpr1a mosaics could
result from a conversion of surface to neural ectoderm fates,
or from an over-proliferation of neural cells. We determined
the proliferating cells per unit area of neural tissue in
Bmpr1a mosaic and wild-type embryos. The number of
proliferating cells in the anterior ectoderm was not signif-
icantly different between Bmpr1a mosaic and wild-type
embryos (P value = 0.68). Overall, our data suggest that
the expanded neural domain results from a conversion of
surface to neural ectoderm fate specifically in the anterior of
the embryo.
Fig. 4. Distribution of neural and surface ectoderm in Bmpr1a mosaics. (A–F) Sox2 whole-mount in situ hybridization and derivative histological sections. (A)
Wild-type, E9.5 embryo, lateral view. Sox2 expression occurs throughout the neural tube. Transverse lines show level of sections in panels B and C. (B and C)
Expression of Sox2 is observed only in neural ectoderm (arrows) at both axial levels. (D) Bmpr1a mosaic E9.5 embryo, ventral view. Levels of sections shown
in E and F indicated by lines. (E) In rostral sections of Bmpr1amosaic embryos, Sox2 expression occurs throughout the columnar ectodermal epithelium, which
is highly convoluted and extends to the margins of the embryo (arrows). Expression is highest medially and tapers to lower levels laterally. (F) In trunk-level
sections, Sox2 expression occurs in the columnar epithelium (arrow); a discrete neural groove (ng) is formed and the columnar epithelium is flanked by low
cuboidal epithelium that does not express Sox2. (G–L) Wnt6 in situ hybridization of wild-type and Bmpr1a mosaic embryos. (G) Wild-type E9.5 embryo,
lateral view. Wnt6 is expressed in surface ectoderm. (H and I) Transverse sections of wild-type embryo hybridized with Wnt6 probe. Surface ectoderm
specifically expressesWnt6 (arrowheads), while neural ectoderm does not expressWnt6. (J) Bmpr1a mosaic, E9.5 embryo, dorsal view. Wnt6 expression in the
amnion (am) occurs as a dark ring at the lateral margins of the embryo. Surface ectoderm expression of Wnt6 is seen caudally (arrowhead). (K) Rostral sections
of mutants show no clearWnt6 expression in ectoderm, only amnion expression (am). (L) Caudally, Wnt6 is expressed in surface ectoderm (arrowheads) lateral
to the thickened medial (neural) ectoderm. (M and N) Eosin-stained transverse sections of E8.5 Bmpr1a mosaic embryo. (M) Rostral section reveals highly
convoluted neural epithelium (arrow), with no histologically detectable surface ectoderm. (N) Trunk section reveals surface ectoderm (arrowhead) lateral to
neural epithelium. Dashed lines indicate transition from neural to surface ectoderm. (O) Area of anterior tissues in wild-type and Bmpr1a mosaic embryos
(E8.0). For each tissue, solid colored bars (left) represent wild-type embryos, and stippled bars (right) represent mosaic embryos. P values are neural = 1.1 �10� 8, mesoderm = 2.4 � 10� 7. Asterisk represents that no morphological evidence for anterior surface ectoderm was observed for Bmpr1a mosaic embryos.
Error bars are F SEM. (P) X-gal and eosin-stained transverse trunk section of Mox2cre/ + ; Bmpr1aflox/null; RosaR26R/ + embryo. (Q) Higher magnification view
of P as indicated. Unrecombined (pink, wild-type) cells are intermingled with recombined (blue, mutant) cells in both neural (arrow) and surface (arrowhead)
ectoderm. Scale bars = 0.25 mm, except Q = 0.025 mm.
S. Davis et al. / Developmental Biology 270 (2004) 47–63 53
S. Davis et al. / Developmental Biology 270 (2004) 47–6354
Anterior–posterior patterning of the neural plate in
Bmpr1a mosaic embryos
To determine the anterior–posterior (A–P) axial level at
which the neuroepithelium becomes expanded, we exam-
ined several regionally restricted gene expression markers.
At E7.5, the boundary between Otx2 and Gbx2 expression
(Fig. 5A) defines the posterior extent of the midbrain
(Wassarman et al., 1997). As in wild-type embryos, the
caudal aspect of the Gbx2 expression domain in mutants
extends to the posterior primitive streak. In contrast, mosaic
embryos show a greatly enlarged domain of Otx2 compared
to stage-matched wild-type control embryos; a clear bound-
ary forms, but it is shifted toward the tail end of the embryo
(Fig. 5B). This indicates that the presumptive brain is
enlarged from early stages.
To further define the A–P axis in Bmpr1a mosaic
embryos, we examined regional neural markers at E8.0.
Krox20 expression marks rhombomeres 3 and 5 (r3 and 5)
in the hindbrain (Fig. 5C) (Swiatek and Gridley, 1993).
Though sometimes obscured by anterior convolutions, both
the r3 and r5 stripes occur in mosaics (Fig. 5D). Caudal to
the r5 stripe, the ectoderm appears relatively normal,
suggesting that the neural epithelium is expanded from
the rostral hindbrain through its anterior limit. We also
noticed that the space between the r3 and r5 stripes
sometimes appears compressed. A similar result was
obtained with the midbrain marker Wnt1 (Wilkinson et
al., 1987); a much narrower band was observed in mutant
embryos relative to wild-type littermates (Figs. 5E and F).
Fig. 5. Expansion of anterior tissues in Bmpr1a mosaics. (A–B) Otx2 (purple)
hindbrain boundary, n indicates node. (A) Wild-type E7.5 embryo, lateral view
hybridization as indicated by arrows. (C) Wild-type E8.5 embryo, dorsal view
hybridization as indicated by arrows. (E) Wild-type E8.5 embryo, lateral vie
hybridization. Dashed lines represent distance between Six3 expression and the
ventral view. Scale bars = 0.5 mm.
In contrast, expression of Six3, which marks the early
forebrain, is expanded in mosaics compared to wild type
(Figs. 5G and H). Intriguingly, sometimes the most rostral
portion of the anterior neural ectoderm showed less intense
staining for forebrain markers than tissue immediately
caudal; we believe this reflects an underlying prechordal
plate defect (see below). Taken together, our results dem-
onstrate that A–P patterning occurs in the neural ectoderm
of Bmpr1a mosaic embryos. A convoluted, broad neural
epithelium extends from the caudal hindbrain forward to the
rostral end of the embryo. Most of this tissue appears to be
of forebrain character, possibly at the expense of the
midbrain and hindbrain.
Dorsal–ventral patterning of the neural plate in Bmpr1a
mosaic embryos
Embryological manipulations and genetic analysis in
zebrafish, Xenopus, and chick suggest that BMP2/4 signal-
ing has a major role in dorsal–ventral (D–V) patterning of
the body plan, including the ectoderm (Lee and Jessell,
1999). We examined D–V patterning in Bmpr1a mosaics to
elucidate the role of BMP2/4 signal transduction during this
process in mouse. Surprisingly, even in very strong mosaics,
the ventral neural tube marker Netrin-1 (Kennedy et al.,
1994) is expressed in domains similarly delimited as in wild
type (Figs. 6A–E). Where the large convolutions of the
expanded anterior neural plate approach the midline, mul-
tiple domains of Netrin-1 expression occur (Fig. 6D). These
data complement and support similar results with Shh
and Gbx2 (light blue) in situ hybridization. Dashed line indicates mid–
(B) Bmpr1a mosaic E7.5 embryo, lateral view. (C–D) Krox20 in situ
. (D) Bmpr1a mosaic E8.5 embryo, ventral view. (E–F) Wnt1 in situ
w. (F) Bmpr1a mosaic E8.5 embryo, lateral view. (G–H) Six3 in situ
hindbrain. (G) Wild-type E8.5, lateral view. (H) Bmpr1a mosaic embryo,
Fig. 6. Dorsal –ventral patterning of the neural tube. (A–E) Netrin in situ hybridization. (A) Wild-type E9.5 embryo, lateral view. (B) Transverse section of
embryo in A at level indicated. (C) Bmpr1a mosaic E9.5 embryo, ventral view. (D) Transverse section of embryo in C at level indicated. (E) Transverse
section of embryo in C at level indicated. (F–J) Msx1 in situ hybridization. (F) Wild-type E9.5 embryo, lateral view. (G) Transverse section through F at
level indicated. (H) Bmpr1a mosaic E9.5 embryo, dorsal view. (I and J) Transverse section of embryo in H at levels indicated. Arrow in I indicates lateral
neuroectoderm with no Msx1 staining. (K and L) Ap2a (purple) and T (light blue) in situ hybridization. (K) Wild-type E8.5 embryo, lateral view. (L) Bmpr1a
mosaic E8.5 embryo, lateral view. (M and N) Sox 10 in situ hybridization. (M) Wild-type E9.5 embryo, lateral view. (N) Bmpr1a mosaic E9.5 embryo,
ventral view. Scale bars = 0.25 mm.
S. Davis et al. / Developmental Biology 270 (2004) 47–63 55
expression (see below), and suggest that the regulation of
ventralizing pathways is largely unperturbed in these em-
bryos. Dorsal neural markers, such as Msx1 and Pax3, are
expressed in dorsal domains in the caudal neural tube and
are sometimes expanded (Figs. 6F–K; data not shown). In
the convoluted, anterior regions devoid of surface ectoderm,
expression of these dorsal markers occurs in regions of the
neurectoderm away from the ventral midline, but not
uniformly in such regions (Fig. 6I).
We further characterized D–V ectodermal patterning by
evaluating neural crest markers. AP2a is an early marker of
the neural crest competence domain (Mitchell et al., 1991)
and is expressed in mosaic embryos (Figs. 6K and L). The
migratory neural crest cell marker Sox10 (Kuhlbrodt et al.,
1998) is expressed in Bmpr1a mosaics in cells located
peripheral to the neural tube, very similar to the pattern
observed in wild type (Figs. 6M and N). Thus, despite the
severe morphological defects of the neural epithelium in
these BMPRIA-deficient embryos, dorsal fates do not ap-
pear to be reduced nor ventral fates increased.
Anterior axial mesendoderm is expanded in BMPRIA
mosaics
To explore the basis for the expanded anterior neural
epithelium in Bmpr1a mosaic embryos, we assessed the
status of tissues known to influence the early development
of the ectoderm. Axial patterning is conferred in part by the
underlying mesendoderm, which derives from the organizer
(Harland and Gerhart, 1997). We used several markers for
the node and axial midline to examine the nature of axial
tissue in Bmpr1a mosaic embryos. The primitive streak,
node, and notochord are all present (Fig. 7A). The axial
midline is always shorter, and the node and axial mesen-
doderm are usually somewhat broader during early somito-
genesis (Figs. 7B and C). This is especially pronounced at
the anterior end of the mesendoderm, which typically
appears broad and diffuse.
The rostral-most axial mesendoderm is the prechordal
plate, which has a key role in patterning the overlying
ventral neural epithelium of the presumptive forebrain
Fig. 7. Axial tissues in Bmpr1a mosaics. (A) A pair of E8.5 littermates, wild-type on left and mosaic on right, showing Otx2 (light blue) and T (purple) in situ
hybridization. nt = notochord, ps = primitive streak (B– I) Shh in situ hybridization. (B) Wild-type E8.5 embryo, ventral view. (C) Bmpr1a mosaic E8.5
embryo, ventral view. Arrow indicates intense rostral staining. Asterisk indicates tissue anterior to Shh expression. For B and C: ame, axial mesendoerm; hg,
hindgut; n, node. (D) Bmpr1a mosaic E9.5 embryo, ventral view. (E–G) Transverse sections of embryo in D at levels indicated. Arrow indicates floor plate
expression, arrowheads indicate notochord expression. PCP, prechordal plate. (H) Wild-type E9.5 embryo, lateral view. (I) Transverse section of embryo in H at
level indicated. Arrow indicates prechordal plate and arrowhead indicates notochord. (J and K) In situ hybridization for Gsc. (J) Wild-type E7.5 embryo,
anterior view. Arrow indicates prechordal plate staining. (K) Bmpr1a mosaic E7.5 embryo, ventral view. Arrow indicates midline, prechordal plate expression
and arrowhead indicates an ectopic focus of expression. Asterisk indicates tissue anterior to Gsc expression. Scale bars = 0.25 mm.
S. Davis et al. / Developmental Biology 270 (2004) 47–6356
(Anderson et al., 2002; Shimamura and Rubenstein, 1997).
Shh is expressed in prechordal plate, as well as in the floor
plate of the ventral neural tube, the notochord, and the
ventral endoderm of the closing foregut and hindgut (Figs.
7B, H, and I). In mosaic embryos, Shh expression typically
appears more diffuse and expanded laterally at anterior
levels (Figs. 7C and D). Histological sections of such
embryos reveal a broad domain of labeled mesendoderm
underlying the anterior neural plate, suggesting an expan-
sion of the prechordal plate (Fig. 7E). In the convoluted
anterior, multiple bends of the neural epithelium express
Shh adjacent to the notochord (Fig. 7F), while at posterior
levels the normal pattern of axial midline expression is
observed (Fig. 7G). This suggests that floorplate identity is
induced in multiple anterior locations where bends in the
neural ectoderm approach the notochord. Note, however,
that despite the analysis of several markers of axial mesen-
doderm over dozens of mosaic embryos, we never saw a
duplicated or partially duplicated axis.
We further examined the state of prechordal plate using
the markers Goosecoid (Gsc) and Dickkopf1 (Dkk1), which
mark this tissue specifically (Camus et al., 2000; Glinka et
al., 1998). Relative to wild-type littermates (Fig. 7J), mosaic
embryos have diffuse, broad domains of Gsc and Dkk1
expression (Fig. 7K; data not shown). However, prechordal
plate markers usually fail to extend to the rostral limit of the
midline as in wild-type embryos (Figs. 7B vs. C); thus, there
is often a portion of apparent neural epithelium that is
anterior to the axial mesendoderm in Bmpr1a mosaics
(asterisks in Figs. 7C and K). Note that this same domain
often expresses forebrain markers less intensely than tissues
immediately to the posterior (Fig. 5F), which likely overlie
the broad prechordal plate. Thus, the axial mesendoderm in
Bmpr1a mosaics is often somewhat broadened but never-
theless quite discrete at trunk levels; moreover, it is active as
an inductive signaling source, as judged by its ability to
induce floor plate character in adjacent neural epithelium. At
the anterior end of the midline mesendoderm, however, the
prechordal plate is expanded and diffuse, underlying an
enlarged portion of the anterior neural plate.
Abnormal distribution of definitive endoderm and anterior
visceral endoderm
The broad prechordal plate in Bmpr1a mosaic embryos
raises the issue of endodermal development in these
S. Davis et al. / Developmental Biology 270 (2004) 47–63 57
mutants, in that the mouse prechordal plate coincides with
the dorsal midline of the foregut endoderm (Shimamura and
Rubenstein, 1997; Sulik et al., 1994). Moreover, the ventral
hindgut domain of Shh is missing in mutant embryos (Figs.
7B and C). We examined the general state of endoderm
using Foxa1 as a marker (Figs. 8A–D), which is expressed
in the endodermal gut tube as well as the overlying
notochord and floor plate (Sasaki and Hogan, 1993). Foxa1
is expressed throughout the ventral surface of Bmpr1a
mosaic embryos (Fig. 8C). Histological sections demon-
strate that the endoderm covers the ventral surface of these
embryos, with no formation of a gut tube (Fig. 8D). Our
analysis of the distribution of recombined cells in mosaic
Fig. 8. Analysis of endoderm via in situ hybridization in Bmpr1a mosaics. (A–D)
plane of section shown in panel B is indicated by the line. (B) Transverse sectio
endodermal gut tube (arrows). Floorplate or notochord expression is indicated by
section shown in panel D is indicated by the line. (D) Transverse section of m
illustrating that endoderm is present but fails to form a gut tube. Floorplate or notoc
E6.75 (late-streak stage) embryo, lateral view, shows anterior primitive streak (aps)
because it is discontinuous with the anterior streak expression domain. However, t
time, which also expresses Foxa2. (F) Foxa2 expression in Bmpr1a mosaic E6.75 (
in lateral region. (G) AFP expression in wild-type E7.0 embryo, lateral view, marki
lateral view. Arrow indicates lateral endodermal patches devoid of AFP staining. (I)
is in newly forming definitive endoderm. (J) mCer1 expression in Bmpr1a mosaic
definitive endoderm. (K) Hesx1 expression in wild-type E7.0 embryo, lateral view
domain. (L) Hesx1 expression in Bmpr1amosaic E7.0 embryo. The extent of the A
between arrowheads in L and K). Scale bars = 0.25 mm.
embryos (Fig. 3H) indicates that this tissue is definitive
endoderm (derived from the epiblast). Even in the 2% of
mosaics that form symmetric headfolds, the morphogenetic
movements required to bring the lateral edges of the
endoderm together do not occur and cause a failure in gut
formation (Fig. 3C; data not shown). Thus, there is a general
failure of endodermal morphogenesis in Bmpr1a mosaics;
moreover, specific domains of the endoderm are abnormal,
such as the expanded prechordal plate.
To analyze the initial stages of endoderm formation, we
assessed markers for the early definitive endoderm and the
visceral endoderm. Initially, the epiblast is covered by
visceral endoderm; during gastrulation, it is replaced by
Foxa1 in situ hybridization. (A) Wild-type E9.5 embryo, lateral view. The
n of wild-type embryo reveals strong expression of Foxa1 throughout the
an arrowhead. (C) Bmpr1a mosaic E9.5 embryo, ventral view. The plane of
osaic embryo shows Foxa1 expression throughout the ventral cell layer,
hordal expression is also present (arrowhead). (E) Foxa2 probe in wild-type
and anterior endoderm expression (en). This is most likely AVE expression
he definitive endoderm from the anterior streak is beginning to form at this
late-streak) embryo, lateral view. Arrow indicates patchy ectopic expression
ng visceral endoderm. (H) AFP expression in Bmpr1a mosaic E7.0 embryo,
mCer1 expression in wild-type E7.0 embryo, lateral view. Distal expression
E7.0 embryo, lateral view. Arrow indicates ectopic, posterior expression in
. Expression occurs in the AVE; arrowheads show boundaries of expression
VE is increased, extending more distally than in wild type (compare distance
S. Davis et al. / Developmental Biology 270 (2004) 47–6358
definitive endoderm, most of which arises from the anterior
primitive streak (Dufort et al., 1998; Lawson et al., 1991).
The visceral endoderm is thus displaced proximally such
that it no longer overlies the epiblast. The anterior visceral
endoderm (AVE) transiently overlies the portion of the
epiblast that gives rise to the brain (Quinlan et al., 1995);
many lines of evidence suggest that the AVE has an
important role in promoting forebrain development (Bed-
dington and Robertson, 1998). At gastrulation, Foxa2 is
expressed in the AVE as well as in anterior primitive streak
and the axial mesendodermal cells that emanate from it
(Fig. 8E) (Sasaki and Hogan, 1993). In Bmpr1a mosaics,
Foxa2 expression occurs in these domains, though each is
broader and less distinct; in addition, ectopic patches of
expression often appeared on the lateral surfaces of these
embryos (Fig. 8F). These patches are very likely ectopic
definitive endodermal precursors because they did not
express markers specific to the AVE (see below) or axial
mesoderm. Further evidence suggesting abnormal morpho-
genesis of the early endoderm is the expression pattern of
the alpha-foetoprotein (AFP) gene, a marker for extraem-
bryonic visceral endoderm (Waldrip et al., 1998). In
Bmpr1a mosaics, AFP was expressed in an uneven, patchy
pattern in mutant embryos (Fig. 8H) relative to wild-type
siblings (Fig. 8G). It is noteworthy that patches devoid of
AFP expression (and thus visceral endoderm) were similar
in distribution and size to the patches of ectopic Foxa2
expression (probable definitive endoderm). These data
suggest an abnormal distribution of visceral and definitive
endoderm during gastrulation, and perhaps an intermingling
of these tissues during gastrulation.
To further investigate the nature of the early endoderm in
Bmpr1a mosaics, we assessed expression of several addi-
tional markers. Mouse Cerberus-1 (mCer1) is expressed
initially in the AVE but then transiently in the newly formed
definitive endoderm as it displaces visceral endoderm
(Pearce et al., 1999). Relative to wild type (Fig. 8I), mCer1
expression in the definitive endoderm not only occurred in
its normal domain but also extended more to the posterior
than in wild type (Fig. 8J). Consistently, markers of the
AVE, such as Hesx1 (Thomas et al., 1995), were properly
restricted to the anterior midline but exhibited an expanded
domain that extended distally toward the anterior primitive
streak (Figs. 8K and L). This suggests that AVE is abnor-
mally displaced by the definitive endoderm. Together, these
results indicate that early endoderm morphogenesis is ab-
normal in Bmpr1a mosaics, with abnormal exposure of the
underlying ectoderm to patterning signals that may emanate
from different regions of the endoderm.
Discussion
By ablating BMPRIA specifically in epiblast cells, we
have circumvented the gastrulation block of Bmpr1a null
embryos to reveal the roles of BMP2/4 signaling in the
development of germ layers. All three germ layers form, and
marked cells lacking Bmpr1a contribute to each at levels
comparable to wild-type marked cells. Thus, BMPRIA is
dispensable for endoderm and mesoderm formation despite
their absence in Bmpr1a null embryos. However, each germ
layer is abnormal; in this report, we focus on endoderm and
ectoderm. Although mutant cells are able to contribute to
surface ectoderm, little if any forms in the anterior of mosaic
embryos. Rather, the prospective forebrain is greatly ex-
panded and convoluted. Morphogenesis of both visceral and
definitive endoderm is abnormal, such that domains known
to promote anterior neurectoderm are expanded during the
early patterning of the ectoderm.
The identity of BMP ligands that signal through
BMPRIA to mediate this morphogenesis and patterning is
unclear. Bmp4 and Bmpr1a have essentially identical null
phenotypes (Mishina et al., 1995; Winnier et al., 1995).
Nonetheless, the defects we observe are unlikely to reflect
unique requirements for BMP4 signaling in the epiblast,
because its loss specifically in the epiblast has relatively
mild defects (Fujiwara et al., 2002). In Xenopus, BMP2,
BMP7 can repress neural and promote surface ectoderm fate
(Suzuki et al., 1997). Nevertheless, in mouse, neither Bmp2
nor Bmp7 is expressed in a way that can account for the
defects, nor is either uniquely required for any of the
processes affected in the Bmpr1a mosaics (Zhao, 2003).
The double and triple null embryos have yet to be generated
to assess potential redundancy. It also remains possible that
an unsuspected BMP is key.
Similarities and differences with zebrafish BMP pathway
mutants
The phenotypes of mouse Bmpr1a epiblast mosaic em-
bryos resemble those of zebrafish loss-of-function mutants
for BMP pathway components, though there are some
notable differences. The mouse phenotypes we observed
are most comparable to the stronger phenotypes observed
for mutants in Bmp2b (swirl), Bmp7 (snailhouse), and their
signal transducer Smad5 (somitabun) (Mullins et al., 1996;
reviewed by Hammerschmidt and Mullins, 2002). Lack of
both maternal and zygotic function of the TGFh type I
receptor lost-a-fin results in a very similar phenotype
(Mintzer et al., 2001). In all these zebrafish mutants, the
axial mesendoderm, derived from the organizer, exhibits a
moderate lateral expansion, while the neural ectoderm is
expanded at the expense of nonneural ectoderm. There are
also similarities in dorsoventral neural patterning (discussed
below). As in the fish mutants, the somitic mesoderm is
expanded and the lateral mesoderm reduced in Bmpr1a
epiblast mosaics (S.M., S.D., J.K. and Y.M., in preparation).
However, whereas we observed more severe defects in
the anterior of Bmpr1a mosaic mouse embryos, the zebra-
fish defects are strong posteriorly and rather mild in anterior
regions. For example, the head structures are comparatively
normal, and there is no apparent broadening of the expres-
S. Davis et al. / Developmental Biology 270 (2004) 47–63 59
sion domain of markers for the most anterior axial mesen-
doderm, the prechordal plate (Mullins et al., 1996; Mintzer
et al., 2001). In contrast, rostral structures are extremely
dysmorphic and convoluted in Bmpr1a mosaics, with broad,
diffuse staining of prechordal plate markers, while the trunk
is much less affected (though neurulation fails at all axial
levels). These differences are likely due in part to very
different morphological constraints confronting the two
types of embryo, such as the large yolk of the fish egg or
the cupped shape of the mouse gastrula.
BMPRIA and morphogenesis of the mammalian embryo
Despite wide variability of Mox2-Cre recombination in
Bmpr1a mosaic embryos, over 90% displayed a character-
istic ‘‘strong’’ phenotype of a severely malformed anterior,
and all mosaics had profound and consistent morphogenesis
defects. The strong phenotype was observed in both low and
high percentage mosaics. Moreover, mutant cells were
evenly represented in all tissue layers and structures ana-
lyzed. These results indicate that the morphogenesis defects
of BMPRIA inactivity are largely cell nonautonomous.
All Bmpr1a mosaic embryos displayed a lack of ventral
closure and embryonic turning. These processes are devel-
opmentally coupled, though the causative cellular and
molecular mechanisms are poorly understood (Kaufman
and Bard, 1999). Furin is a protease involved in the
maturation of growth factors, including BMP4 (Constam
and Robertson, 2000), and is required in the definitive
endoderm for both ventral closure and turning (Constam
and Robertson, 2000). Embryos lacking Smad5, encoding a
BMPRIA signal transducer, also show defective ventral
closure and embryonic turning (Chang et al., 1999). Col-
lectively, these results suggest that BMP signaling, acting
through BMPRIA, is required for the coordinate regulation
of these morphogenetic processes.
We also observed abnormal early endodermal morpho-
genesis in Bmpr1a mosaic embryos. During gastrulation,
the definitive endoderm (DE) spreads more to the posterior
in the distal portion of the embryo than is seen in wild type.
The visceral endoderm appeared to be aberrantly displaced
in mosaic embryos. There was some evidence of intermin-
gling of visceral and definitive endoderm in lateral portions
of the embryo. Consistently, we saw that the AVE extended
further distally than in littermates, suggesting that it had not
been displaced as normal by DE from the anterior primitive
streak. This may result from some of the DE occurring
more posteriorly, and thus exerting less of a displacing
effect on AVE. However, by E8.0, the ventral surface of the
embryo was covered by endoderm that was largely if not
exclusively of epiblast origin, that is, definitive endoderm.
(We cannot exclude that some of the unlabeled cells in the
endoderm could be of visceral origin, but they certainly
include unrecombined cells of epiblast origin.) We ob-
served that the node was often mediolaterally expanded,
at least as judged by the expression of node markers.
Accordingly, it is conceivable that the anterior primitive
streak (from which both DE and node derive) might also be
expanded. Thus, BMPRIA signaling may have a direct role
in DE formation or migration. In the mosaics, the DE tissue
layer forms but the timing and distribution of its spread are
abnormal.
The molecular basis of the morphogenesis pathway(s)
downstream of BMPRIA is unclear. Our results demonstrate
that BMP signaling through BMPRIA is required to orches-
trate several aspects of germ layer morphogenesis. Given
the nonautonomy of morphogenetic defects, it is likely that
BMPRIA signal transduction promotes the expression or
activity of a secreted factor or factors that in turn regulate
the cellular behaviors underlying specific morphogenetic
events. In zebrafish, BMP signaling may affect convergent-
extension morphogenesis by regulating the expression of
Wnt genes (Myers et al., 2002), which in turn encode
secreted factors. The individual morphogenetic movements
of early embryogenesis are very poorly understood in the
mouse, but the parallels with zebrafish and other models
provide hypotheses to test.
We speculate that BMPRIA signaling promotes the
production of an unidentified epiblast morphogenetic signal,
but we do not know its specific source or identity. Good
arguments can be made for the candidacy of various Wnt,
FGF, or Nodal ligands, all of which are active in establish-
ing the body plan of the mouse embryo (reviewed by Lu et
al., 2001). We are examining these factors in ongoing
experiments.
BMP signaling and mediolateral regionalization of the
ectoderm
The ‘‘default model’’ of neural induction holds that
naive ectodermal cells that transduce a BMP2/4 signal will
become surface ectoderm, while those that do not trans-
duce this signal will become neural (Chang and Hemmati-
Brivanlou, 1998). Consistent with this model, we observe
little if any surface ectoderm rostral to the hindbrain.
Instead, neurectoderm is expanded, extending to the em-
bryonic margins. However, caudal to the hindbrain,
Bmpr1a mosaic embryos have significant amounts of
surface ectoderm, even in very strong phenotypes (though
probably less than in corresponding wild-type embryos).
This tissue contains cells that have undergone recombina-
tion and, therefore, presumably lack Bmpr1a. These data
suggest that the lack of BMPRIA may inhibit the surface
ectoderm fate while not precluding it. Conversely,
BMPRIA activity may promote surface ectoderm forma-
tion, but is not essential for its formation.
Evidence from many experimental models suggests that
BMP2/4 signaling promotes dorsal fates and inhibits ventral
fates in the nascent neural plate (Lee and Jessell, 1999; Sasai
and De Robertis, 1997). Indeed, zebrafish Bmp2b/swirl
mutants show lack of neural crest, reduced dorsal neural
markers, and an expanded floorplate (Barth et al., 1999;
S. Davis et al. / Developmental Biology 270 (2004) 47–6360
Nguyen et al., 2000). However, in Bmpr1a mosaic mouse
embryos, the D–V axis of the neural tube is surprisingly
well patterned. Although neurulation fails at all levels,
dorsal–lateral neural plate markers are not reduced, and
are in some cases even expanded. Oddly, expanded domains
are uneven and discontinuous, possibly related to the local
degree of mosaicism in the ectoderm itself or in underlying
mesoderm. Even in very strong mosaics, the neural crest is
specified and can migrate away from the neural plate.
Meanwhile, ventral markers are not expanded in general,
with the floorplate being of essentially normal width.
However, ventral regions are sometimes expanded in dys-
morphic anterior regions, particularly where folds of ecto-
derm come near the ventral midline. This indicates that the
ventralizing capacity of the axial midline is intact, but not
necessarily greater than in wild-type, as one might have
expected with reduced BMP signaling.
Despite the absence of what is presumed to be the key
BMP2/4 type I receptor during mouse gastrulation and
neurulation, we observed less severe effects on mediolateral
ectodermal development than expected, given the data from
Xenopus experiments. One possibility is that BMP signaling
is indeed dispensable for surface ectoderm formation and
dorsal neural patterning. Alternatively, we may have failed
to sufficiently diminish BMP signaling. Recombination by
Mox2-Cre may be too late, though it occurs throughout the
embryo well before gastrulation begins (Tallquist and Sor-
iano, 2000); perhaps BMPRIA is a very stable protein that
persists sufficiently long. A more likely possibility is that a
related receptor can transduce any required signals. Bmpr1b
is neither expressed nor required in early mouse develop-
ment (Beppu et al., 2000). Activin receptor 1A (Alk2) can
bind some BMP ligands (Liu et al., 1995; Macias-Silva et
al., 1998). In Xenopus, activated Alk2 can repress neural but
promote surface ectoderm fates (Suzuki et al., 1997). In
mouse, it does not appear to be required in the epiblast for
development of either neural or surface ectoderm lineages
(Gu et al., 1999; Mishina et al., 1999). Thus, Alk2 and
BMPRIA have distinct roles in the epiblast, but may also
function redundantly in neuralization. The absolute role of
BMPRIA signaling in neural pattern formation must be
reexamined when reagents are available to completely
remove it from the epiblast.
Expanded anterior neurectoderm in Bmpr1a mosaics
A striking defect of more than 90% of the Bmpr1a
mosaics we observed is that the neural ectoderm is expand-
ed and convoluted from the hindbrain forward. This appears
to be the expense of anterior surface ectoderm, which
appeared to be absent by virtue of both morphology and a
surface ectoderm gene expression marker. Though the
anterior neurectoderm in mosaics occupies more surface
area than in wild-type counterparts, we saw no increased
proliferation in the mosaic neuroepithelium. These results
suggest that the proportion of ectoderm that is neuralized is
greatly increased in the anterior section of Bmpr1a mosaics.
This is not simply a consequence of these ectodermal cells
being unable to receive BMP signals, as the ‘‘neural
default’’ model would suggest, because the expanded neural
ectoderm was composed largely of wild-type cells in some
mosaics with the expanded anterior neurectoderm pheno-
type (e.g., Fig. 3F).
Regional gene expression markers for A–P neural
domains indicate that an expanded presumptive forebrain
accounts for most of the enlarged anterior neurectoderm of
Bmpr1a mosaics. In fact, other domains, such as midbrain
and hindbrain, may even be reduced along the A–P axis
relative to wild type. We suggest that these observations can
best be explained by an enlarged area of presumptive
anterior ectoderm being exposed to forebrain-promoting
factors early in development.
Primary anterior signaling centers are expanded in Bmpr1a
mosaic embryos
We observed that the prechordal plate (PrCP), residing at
the rostral limit of the axial mesendoderm, was broad and
often diffuses in Bmpr1a mosaic embryos. Exposure of
mouse ectoderm explants to prechordal plate induces fore-
brain markers (Shimamura and Rubenstein, 1997). Because
the mosaic PrCP extends laterally, it underlies a broader
region of ectoderm and may influence prospective surface
ectoderm toward an anterior neural fate. Consistent with this
view, we observed that the PrCP sometimes failed to extend
to the rostral extreme of the neurectoderm in Bmpr1a
mosaics, exactly the region of anterior neurectoderm that
shows weaker expression of forebrain markers such as Six3.
We have found that increased BMP signaling is detrimental
to the PrCP and forebrain development in mouse (Anderson
et al., 2002). Thus, it is consistent that decreased BMP
signal reception in Bmpr1a mosaics seems to enhance the
influence of the widened PrCP.
We also observed that in virtually all Bmpr1a mosaic
embryos examined at early to mid-gastrulation, the anterior
visceral endoderm (AVE) was expanded. In particular, the
AVE occupied a greater proportion of the A–P axis of the
mosaic embryos. Increasing evidence supports the hypoth-
esis that AVE initializes the forebrain fate (Beddington and
Robertson, 1998), likely via repression of posteriorizing
factors in the naive neurectoderm (Perea-Gomez et al.,
2001). However, the AVE is not sufficient to induce and
maintain prospective forebrain, but rather requires the
synergy of organizer derivatives (Tam and Steiner, 1999).
In mouse, BMP activity appears to be inhibitory to both
AVE and organizer function (Bachiller et al., 2000), and a
key function of the organizer (Harland and Gerhart, 1997)
and possibly the AVE (Beddington and Robertson, 1998) is
to inhibit BMP signaling. Therefore, with the reduced
overall BMP signaling in Bmpr1a mosaics, the epiblast
might be sensitized to the anteriorizing influence of both
the AVE and the organizer.
S. Davis et al. / Developmental Biology 270 (2004) 47–63 61
Our data fit well with the ‘‘double assurance model’’ for
anterior neural patterning in the mouse (Shawlot et al.,
1999; Thomas and Beddington, 1996). In this scheme,
forebrain fate is promoted by the AVE, but is reinforced
and maintained by the anterior axial mesendoderm (e.g.,
PrCP). Consistent with the general point of the neural
default model for initial neuralization (that reduced BMP
signaling promotes a basal neural fate of anterior charac-
ter), the ectoderm may be sensitized to both of these
anteriorizing influences by a reduced overall level of
BMP signaling in the Bmpr1a mosaic embryos. Bmpr1a
mosaics have expanded AVE and PrCP, probably as a
result of abnormal morphogenesis of visceral and defini-
tive endoderm. We suggest that the expanded AVE overlies
a greater extent of epiblast and induces it to a forebrain
fate, thus increasing the proportion of labile forebrain in
the nascent neurectoderm. Subsequently, the broad PrCP
stabilizes forebrain identity in the neural ectoderm in its
vicinity.
Acknowledgments
We thank B. Hogan, E. Meyers, and members of the
Klingensmith lab for helpful comments on the manuscript.
Plasmids were kindly provided by A. McMahon, S.
Tilghman, P. Mitchell, B. Hogan. M. Tessier-Lavigne, T.
Gridley, and R. Lovell-Badge. We are grateful to P. Soriano
and M. Tallquist for providing Mox2-Cre mice. This work
was supported by NIH awards to J.K. (R01DE013674 and
P01HD39948).
References
Anderson, R.M., Lawrence, A.R., Stottmann, R.W., Bachiller, D., Klingen-
smith, J., 2002. Chordin and noggin promote organizing centers of
forebrain development in the mouse. Development 129, 4975–4987.
Ang, S.L., Conlon, R.A., Jin, O., Rossant, J., 1994. Positive and negative
signals from mesoderm regulate the expression of mouse Otx2 in ecto-
derm explants. Development 120, 2979–2989.
Bachiller, D., Klingensmith, J., Kemp, C., Belo, J.A., Anderson, R.M.,
May, S.R., McMahon, J.A., McMahon, A.P., Harland, R.M., Rossant,
J., et al., 2000. The organizer factors Chordin and Noggin are required
for mouse forebrain development. Nature 403, 658–661.
Barth, K.A., Kishimoto, Y., Rohr, K.B., Seydler, C., Schulte-Merker, S.,
Wilson, S.W., 1999. Bmp activity establishes a gradient of positional
information throughout the entire neural plate. Development 126,
4977–4987.
Beddington, R.S., Robertson, E.J., 1998. Anterior patterning in the mouse.
Trends Genet. 14, 277–284.
Belo, J.A., Bouwmeester, T., Leyns, L., Kertesz, N., Gallo, M., Follettie,
M., De Robertis, E.M., 1997. Cerberus-like is a secreted factor with
neutralizing activity expressed in the anterior primitive endoderm of
the mouse gastrula. Mech. Dev. 68, 45–57.
Beppu, H., Kawabata, M., Hamamoto, T., Chytil, A., Minowa, O., Noda,
T., Miyazono, K., 2000. BMP type II receptor is required for gastru-
lation and early development of mouse embryos. Dev. Biol. 221,
249–258.
Camus, A., Davidson, B.P., Billiards, S., Khoo, P., Rivera-Perez, J.A.,
Wakamiya, M., Behringer, R.R., Tam, P.P., 2000. The morphogenetic
role of midline mesendoderm and ectoderm in the development of the
forebrain and the midbrain of the mouse embryo. Development 127,
1799–1813.
Chang, C., Hemmati-Brivanlou, A., 1998. Cell fate determination in em-
bryonic ectoderm. J. Neurobiol. 36, 128–151.
Chang, H., Huylebroeck, D., Verschueren, K., Guo, Q., Matzuk, M.M.,
Zwijsen, A., 1999. Smad5 knockout mice die at mid-gestation due to
multiple embryonic and extraembryonic defects. Development 126,
1631–1642.
Constam, D.B., Robertson, E.J., 2000. Tissue-specific requirements for the
proprotein convertase furin/SPC1 during embryonic turning and heart
looping. Development 127, 245–254.
Dewulf, N., Verschueren, K., Lonnoy, O., Moren, A., Grimsby, S., Vande
Spiegle, K., Miyazono, K., Huylebroeck, D., ten Dijke, P., 1995. Dis-
tinct spatial and temporal expression patterns of two type I receptors for
Bone Morphogenetic Proteins during mouse embryogenesis. Endocri-
nology 136, 2652–2663.
Downs, K.M., Davies, T., 1993. Staging of gastrulating mouse embryos by
morphological landmarks in the dissecting microscope. Development
118, 1255–1266.
Dufort, D., Schwartz, L., Harpal, K., Rossant, J., 1998. The transcription
factor HNF3beta is required in visceral endoderm for normal primitive
streak morphogenesis. Development 125, 3015–3025.
Echelard, Y., Epstein, D.J., St-Jacques, B., Shen, L., Mohler, J., McMahon,
J.A., McMahon, A.P., 1993. Sonic hedgehog, a member of a family of
putative signaling molecules, is implicated in the regulation of CNS
polarity. Cell 75, 1417–1430.
Fainsod, A., Steinbeisser, H., De Robertis, E.M., 1994. On the function of
BMP-4 in patterning the marginal zone of the Xenopus embryo. EMBO
J. 13, 5015–5025.
Fujiwara, T., Dehart, D.B., Sulik, K.K., Hogan, B.L., 2002. Distinct
requirements for extra-embryonic and embryonic bone morphogenetic
protein 4 in the formation of the node and primitive streak and coor-
dination of left – right asymmetry in the mouse. Development 129,
4685–4696.
Gavin, B.J., McMahon, J.A., McMahon, A.P., 1990. Expression of multiple
novel Wnt-1/int-1-related genes during fetal and adult mouse develop-
ment. Genes Dev. 4, 2319–2332.
Glinka, A., Wu, W., Onichtchouk, D., Blumenstock, C., Niehrs, C., 1997.
Head induction by simultaneous repression of Bmp and Wnt signalling
in Xenopus. Nature 389, 517–519.
Glinka, A., Wu, W., Delius, H., Monaghan, A.P., Blumenstock, C., Niehrs,
C., 1998. Dickkopf-1 is a member of a new family of secreted proteins
and functions in head induction. Nature 391, 357–362.
Gu, Z., Reynolds, E.M., Song, J., Lei, H., Feijen, A., Yu, L., He, W.,
MacLaughlin, D.T., van den Eijnden-van Raaij, J., Donahoe, P.K.,
et al., 1999. The type I serine/threonine kinase receptor ActRIA
(ALK2) is required for gastrulation of the mouse embryo. Development
126, 2551–2561.
Hammerschmidt, M., Mullins, M.C., 2002. Dorsoventral patterning in the
zebrafish: bone morphogenetic proteins and beyond. Results Probl. Cell
Differ. 40, 72–95.
Harland, R., Gerhart, J., 1997. Formation and function of Spemann’s or-
ganizer. Annu. Rev. Cell Dev. Biol. 13, 611–667.
Hartley, K.O., Hardcastle, Z., Friday, R.V., Amaya, E., Papalopulu, N.,
2001. Transgenic Xenopus embryos reveal that anterior neural develop-
ment requires continued suppression of BMP signaling after gastrula-
tion. Dev. Biol. 238, 168–184.
Hayashi, S., Lewis, P., Pevny, L., McMahon, A.P., 2002. Efficient gene
modulation in mouse epiblast using a Sox2Cre transgenic mouse strain.
Gene Expr. Patterns 2, 93–97.
Hogan, B.L., Beddington, R., Costantini, F., Lacy, E., 1994. Manipulating
the Mouse Embryo. Cold Spring Harbor Press, New York.
Kaufman, M.H., 1992. The Atlas of Mouse Development. Academic Press,
Inc., San Diego.
S. Davis et al. / Developmental Biology 270 (2004) 47–6362
Kaufman, M.H., Bard, J.B.L., 1999. The Anatomical Basis of Mouse De-
velopment. Academic Press, Inc., San Diego.
Kennedy, T.E., Serafini, T., de la Torre, J.R., Tessier-Lavigne, M., 1994.
Netrins are diffusible chemotropic factors for commissural axons in the
embryonic spinal cord. Cell 78, 425–435.
Knecht, A.K., Bronner-Fraser, M., 2002. Induction of the neural crest: a
multigene process. Nat. Rev. Genet. 3, 453–461.
Kuhlbrodt, K., Herbarth, B., Sock, E., Hermans-Borgmeyer, I., Wegner, M.,
1998. Sox10, a novel transcriptional modulator in glial cells. J. Neurosci.
18, 237–250.
Lawson, K.A., Meneses, J.J., Pedersen, R.A., 1991. Clonal analysis of
epiblast fate during germ layer formation in the mouse embryo. Devel-
opment 113, 891–911.
Lee, K.J., Jessell, T.M., 1999. The specification of dorsal cell fates in
the vertebrate central nervous system. Annu. Rev. Neurosci. 22,
261–294.
Liu, F., Ventura, F., Doody, J., Massague, J., 1995. Human type II receptor
for bone morphogenic proteins (BMPs): extension of the two-kinase
receptor model to the BMPs. Mol. Cell Biol. 15, 3479–3486.
Lu, C.C., Brennan, J., Robertson, E.J., 2001. From fertilization to gastru-
lation: axis formation in the mouse embryo. Curr. Opin. Genet. Dev. 11,
384–392.
Macias-Silva, M., Hoodless, P.A., Tang, S.J., Buchwald, M., Wrana, J.L.,
1998. Specific activation of Smad1 signaling pathways by the BMP7
type I receptor, ALK2. J. Biol. Chem. 273, 25628–25636.
Mackenzie, A., Ferguson, M.W.J., Sharpe, P.T., 1991. Hox-7 expression
during murine craniofacial development. Development 113, 601–611.
Massague, J., Chen, Y.G., 2000. Controlling TGF-beta signaling. Genes
Dev. 14, 627–644.
Mayor, R., Aybar, M.J., 2001. Induction and development of neural crest in
Xenopus laevis. Cell Tissue Res. 305, 203–209.
McMahon, J.A., Takada, S., Zimmerman, L.B., Fan, C.M., Harland, R.M.,
McMahon, A.P., 1998. Noggin-mediated antagonism of BMP signaling
is required for growth and patterning of the neural tube somite. Genes
Dev. 12, 1438–1452.
Mintzer, K.A., Lee, M.A., Runke, G., Trout, J., Whitman, M., Mullins,
M.C., 2001. Lost-a-fin encodes a type I BMP receptor, Alk8, acting
maternally and zygotically in dorsoventral pattern formation. Develop-
ment 128, 859–869.
Mishina, Y., Suzuki, A., Ueno, N., Behringer, R.R., 1995. Bmpr encodes a
type I bone morphogenetic protein receptor that is essential for gastru-
lation during mouse embryogenesis. Genes Dev. 9, 3027–3037.
Mishina, Y., Crombie, R., Bradley, A., Behringer, R.R., 1999. Multiple
roles for activin-like kinase-2 signaling during mouse embryogenesis.
Dev. Biol. 213, 314–326.
Mishina, Y., Hanks, M.C., Miura, S., Tallquist, M.D., Behringer, R.R.,
2002. Generation of Bmpr/Alk3 conditional knockout mice. Genesis
32, 69–72.
Mitchell, P.J., Timmons, P.M., Hebert, J.M., Rigby, P.W., Tijan, R., 1991.
Transcription factor AP-2 is expressed in neural crest cell lineages
during mouse embryogenesis. Genes Dev. 5, 105–119.
Miyazono, K., Kusanagi, K., Inoue, H., 2001. Divergence and convergence
of TGF-beta/BMP signaling. J. Cell. Physiol. 187, 265–276.
Mullins, M.C., Hammerschmidt, M., Kane, D.A., Odenthal, J., Brand, M.,
van Eeden, F.J., Furutani-Seiki, M., Granato, M., Haffter, P., Heisen-
berg, C.P., Jiang, Y.J., Kelsh, R.N., Nusslein-Volhard, C., 1996. Genes
establishing dorsoventral pattern formation in the zebrafish embryo: the
ventral specifying genes. Development 123, 81–93.
Myers, D.C., Sepich, D.S., Solnica-Krezel, L., 2002. Bmp activity gradient
regulates convergent extension during zebrafish gastrulation. Dev. Biol.
243, 81–98.
Nguyen, V.H., Trout, J., Connors, S.A., Andermann, P., Weinberg, E.,
Mullins, M.C., 2000. Dorsal and intermediate neuronal cell types of
the spinal cord are established by a BMP signaling pathway. Develop-
ment 127, 1209–1220.
Oliver, G., Mailhos, A., Wehr, R., Copeland, N.G., Jenkins, N.A., Gruss, P.,
1995. Six3, a murine homologue of the sine oculis gene, demarcates the
most anterior border of the developing neural plate and is expressed
during eye development. Development 121, 4045–4055.
Pearce, J.J., Penny, G., Rossant, J., 1999. A mouse cerberus/Dan-related
gene family. Dev. Biol. 209, 98–110.
Perea-Gomez, A., Rhinn, M., Ang, S.L., 2001. Role of the anterior visceral
endoderm in restricting posterior signals in the mouse embryo. Int. J.
Dev. Biol. 45, 311–320.
Quinlan, G.A., Williams, E.A., Tan, S.S., Tam, P.P., 1995. Neuroectoder-
mal fate of epiblast cells in the distal region of the mouse egg cylinder:
implication for body plan organization during early embryogenesis.
Development 121, 87–98.
Sasai, Y., De Robertis, E.M., 1997. Ectodermal patterning in vertebrate
embryos. Dev. Biol. 182, 5–20.
Sasaki, H., Hogan, B.L., 1993. Differential expression of multiple fork
head related genes during gastrulation and axial pattern formation in
the mouse embryo. Development 118, 47–59.
Schultheiss, T.M., Burch, J.B., Lassar, A.B., 1997. A role for bone mor-
phogenetic proteins in the induction of cardiac myogenesis. Genes Dev.
11, 451–462.
Shawlot, W., Wakamiya, M., Kwan, K.M., Kania, A., Jessell, T.M., Beh-
ringer, R.R., 1999. Lim1 is required in both primitive streak-derived
tissues and visceral endoderm for head formation in the mouse. Devel-
opment 126, 4925–4932.
Shimamura, K., Rubenstein, J.L., 1997. Inductive interactions direct early
regionalization of the mouse forebrain. Development 124, 2709–2718.
Soriano, P., 1999. Generalized lacZ expression with the ROSA26 Cre
reporter strain. Nat. Genet. 21, 70–71.
Sulik, K., Dehart, D.B., Iangaki, T., Carson, J.L., Vrablic, T., Gesteland, K.,
Schoenwolf, G.C., 1994. Morphogenesis of the murine node and noto-
chordal plate. Dev. Dyn. 201, 260–278.
Suzuki, A., Ueno, N., Hemmati-Brivanlou, A., 1997. Xenopus msx1 medi-
ates epidermal induction and neural inhibition by BMP4. Development
124, 3037–3044.
Swiatek, P.J., Gridley, T., 1993. Perinatal lethality and defects in hindbrain
development in mice homozygous for a targeted mutation of the zinc
finger gene Krox20. Genes Dev. 7, 2071–2084.
Tallquist, M.D., Soriano, P., 2000. Epiblast-restricted Cre expression in
MORE mice: a tool to distinguish embryonic vs. extra-embryonic gene
function. Genesis 26, 113–115.
Tam, P.P., Steiner, K.A., 1999. Anterior patterning by synergistic activity of
the early gastrula organizer and the anterior germ layer tissues of the
mouse embryo. Development 126, 5171–5179.
Thomas, P., Beddington, R., 1996. Anterior primitive endoderm may be
responsible for patterning the anterior neural plate in the mouse embryo.
Curr. Biol. 6, 1487–1496.
Thomas, P.Q., Johnson, B.V., Rathjen, J., Rathjen, P.D., 1995. Sequence,
genomic organization, and expression of the novel homeobox gene
Hesx1. J. Biol. Chem. 270, 3869–3875.
Tiso, N., Filippi, A., Pauls, S., Bortolussi, M., Argenton, F., 2002. BMP
signalling regulates anteroposterior endoderm patterning in zebrafish.
Mech. Dev. 118, 29–37.
Waldrip, W.R., Bikoff, E.K., Hoodless, P.A., Wrana, J.L., Robertson, E.J.,
1998. Smad2 signaling in extraembryonic tissues determines anterior–
posterior polarity of the early mouse embryo. Cell 92, 797–808.
Wassarman, K.M., Lewandoski, M., Campbell, K., Joyner, A.L., Ruben-
stein, J.L.R., Martinez, S., Martin, G.R., 1997. Specification of the an-
terior hindbrain and establishment of a normal mid/hindbrain organizer
is dependent on Gbx2 gene function. Development 124, 2923–2934.
Weinstein, D.C., Hemmati-Brivanlou, A., 1999. Neural induction. Annu.
Rev. Cell Dev. Biol. 15, 411–433.
Wilkinson, D.G., Bailes, J.A., McMahon, A.P., 1987. Expression of the
proto-oncogene int-1 is restricted to specific neural cells in the devel-
oping mouse embryo. Cell 50, 79–88.
Wilkinson, D.G., Bhatt, S., Herrmann, B.G., 1990. Expression pattern of
the mouse T gene and its role in mesoderm formation. Nature 343,
657–659.
Winnier, G., Blessing, M., Labosky, P.A., Hogan, B.L.M., 1995. Bone
S. Davis et al. / Developmental Biology 270 (2004) 47–63 63
morphogenetic protein-4 is required for mesoderm formation and pat-
terning in the mouse. Genes Dev. 9, 2105–2116.
Wood, H.B., Episkopou, V., 1999. Comparative expression of the mouse
Sox1, Sox2 and Sox3 genes from pre-gastrulation to early somite
stages. Mech. Dev. 86, 197–201.
Wu, L., de Bruin, A., Saavedra, H.I., Starovic, M., Trimboli, A., Yang, Y.,
Opavska, J., Wilson, P., Thompson, J.C., Ostrowski, M.C., et al., 2003.
Extra-embryonic function of Rb is essential for embryonic development
and viability. Nature 421, 942–947.
Yi, S.E., Daluiski, A., Pederson, R., Rosen, V., Lyons, K.M., 2000. The
type I BMP receptor BMPRIB is required for chondrogenesis in the
mouse limb. Development 127, 621–630.
Zhao, G.Q., 2003. Consequences of knocking out BMP signaling in the
mouse. Genesis 35, 43–56.